kinds of scientific rationalism: the case for ... · kinds of scientific rationalism: the case for...

Kinds of Scientific Rationalism: The Case for

Methodological Liberalism

by

Scott Scheall

CHOPE Working Paper No. 2015-05

May 2015

Kinds of Scientific Rationalism: The Case for Methodological Liberalism Scott Scheall Draft as of 5/21/2015. Please do not quote or circulate.

1

Abstract:

The present paper considers the implications of the postulate that the activities of scientists

constitute complex phenomena in the sense associated with the methodological writings of the

Nobel Prize-winning Austrian economist, methodologist, and political philosopher, F.A. Hayek.

Although Hayek wrote extensively on the methodology of sciences that investigate systems of

complex phenomena, he never addressed the possibility that science itself is such a system. The

application of Hayek’s methodology of sciences of complex phenomena to science itself implies

some minimal criteria for explanations of scientific rationality. If science is complex in Hayek’s

sense, then scientific belief may be rational in more than one way. It is argued that a failure to

recognize the possibility of multiple kinds of scientific rationality contributes to an error theory

of certain unsuccessful accounts of scientific belief in the history of philosophy of science. It is

further argued that, where ecological rationality is operative, rational belief requires an element

of methodological liberty. It is shown that acceptance of the possibility of ecologically-rational

scientific outcomes – a view here dubbed methodological liberalism – is closely related to

Hayek's denial of the possibility of a successful scientism, a denial crucial to his arguments

against socialism and Keynesian macroeconomics.

Keywords: F.A. Hayek, philosophy of science, complex phenomena, ecological rationality,

methodological liberalism

JEL Codes: B29, B31, B41, P11


2

There seems to me to exist a sort of rationalism which, by not recognizing [the] limits of

the powers of individual reason, in fact tends to make human reason a less effective

instrument than it could be…[T]he best name for this kind of naïve rationalism is

rationalist constructivism. It is a view which in the social sphere has…wrought

immeasurable harm [.] Rationalism in this sense is the doctrine which assumes that…we

have it in our power so to shape our institutions that of all possible sets of results that

which we prefer to all others will be realized; and that our reason should never resort to

automatic or mechanical devices when conscious consideration of all factors would make

preferable an outcome different from that of the spontaneous process.

–F.A. Hayek, “Kinds of Rationalism” ([1964a] 2014, 41–42)

1. Introduction

The present paper considers the implications of the postulate that the activities of scientists (qua

scientists) constitute complex phenomena in the sense associated with the methodological

writings of the Nobel Prize-winning Austrian economist,1 methodologist, and political

philosopher, F.A. Hayek. Although Hayek wrote extensively on the methodology of sciences that

investigate systems of complex phenomena, he never explicitly addressed the possibility that

1 Hayek was Austrian both by birth and by economic-scientific predilection. That is, he was both born in Austria (Vienna, to be exact) and a prominent member of the Austrian School of economics.


3

science itself is such a system.2 An analysis of science’s possible complexity, such as occupies

the first two parts of the present paper, leads to a number of interesting conclusions.

First, it is argued that the application of Hayek’s methodology of sciences of complex

phenomena to science itself implies some minimal criteria for explanations of science, i.e., a

meta-methodology.3 For Hayek, complex phenomena are those with respect to which our

knowledge is necessarily deficient: we lack either an adequate theory of the phenomena or data

sufficient to generate precise explanations and predictions (these being two sides of the same

coin on Hayek’s epistemology) of events in the relevant domain. Thus, if science is complex in

Hayek’s sense, then our ability to explain science is necessarily limited. In particular, scientific

beliefs – or, at least, some scientific beliefs – might be rational, yet we would be unable to fully

account for this rationality. To adopt a locution of one of Hayek’s followers, Vernon Smith (also

a Nobel Laureate in Economics), such beliefs would then be rational in an “ecological” sense4

2 However, a number of modern Hayekians have developed a model of science as an “adaptive classifying system” (see McQuade and Butos (2005), McQuade (2007), McQuade (2010)). The reader interested in such a model is referred to these essays. The present paper is not concerned with the development and defense of a particular model of the phenomena of scientific activity as complex, but with the general implications for methodology of science’s possible complexity. These authors explain science on the model of a particular kind of emergent order, namely, Hayek’s (1952) model of an organism’s “sensory order,” while I consider science from the perspective of Hayek’s more general theory of complex phenomena. Their analysis entails mine, but not the other way around: the phenomena of scientific activity cannot be like the sensory order without being complex, but may be complex without being like the sensory order. 3 In this respect, the current essay bears some resemblance to Butos and Koppl (2003) and McQuade (2010). However, the argument of the present essay is sufficiently unique to warrant a distinct treatment. It should also be mentioned that the meta-methodological implications of a model of scientific phenomena as complex in Hayek’s sense are not dissimilar to those of other treatments of science as an evolutionary process (such as, e.g., Toulmin (1981), Popper (1984), and Hull (1988)). 4 Smith’s (2008, 2) discussion of the distinction between the constructivist and ecological varieties of rationality is worth quoting at length:

Constructivist rationality, applied to individuals or organizations, involves the deliberate use of reason to analyze and prescribe actions judged to be better than alternative feasible actions that might be chosen. When applied to institutions, constructivism involves the deliberate design of rule systems to achieve desirable performance…


4

and no methodology would be tenable that failed to recognize the possibility of ecologically-

rational scientific beliefs.

Second, that they failed to assume this possibility within the context of justification, it is

argued, contributes to an explanation of the perceived shortcomings of certain unsuccessful

accounts of scientific belief. The history of philosophy of science is, at least in part, the story of

the search for an explicit explanation of the scientific method’s apparent superior rationality as a

mode of belief fixation. Of course, if there are scientific beliefs that are ecologically rational and

which, therefore, cannot be fitted into a constructivist framework, then this project was bound to

fail. Moreover, those who reacted to the failure of such constructivist philosophies of science by

simply denying, against all appearances, the relevant datum of science’s superior rationality

made the same error of assuming that, if it exists, this rationality must be completely effable.

Thus, a Hayekian meta-methodology contributes to an error theory of these failed doctrines,

namely, the error of failing to recognize a third possibility other than constructivist rationalism

and irrationalism.

Third, it follows that coercion cannot be an effective method where scientific beliefs are

rational in an ecological sense. And this is so for purely epistemic reasons, without regard to

ethical qualms or questions of the incentive structure confronting coerced scientists. If it is not

Ecological rationality refers to emergent orders in the form of practices, norms, and evolving institutional rules governing action by individuals that are part of our cultural and biological heritage and are created by human interactions, but not by conscious design. The two concepts are not inherently in opposition; the issues are emphatically not about constructivist versus ecological rationality…and in fact the two can and do work together. For example, in evolutionary processes, constructivist cultural innovations can provide variations while ecological fitness processes do the work of selection. [italics in the original]

Like Smith, Hayek ([1964a] 2014, 43) recognized that the kinds of rationality “constantly interact.” Butos and Koppl (2003) coin the phrase “processive” rationality to describe the kind of rationality involved in the emergence of complex orders.


5

fully understood how to arrive at a rational scientific belief, then, regardless of the coercive

techniques adopted, because the coercer doesn’t possess the knowledge required to coerce

effectively, one cannot be coerced into believing rationally. It further follows that an element of

liberty is necessary (but, of course, by no means sufficient) for the realization of rational

scientific belief in such cases.

Fourth and finally, it is argued that Hayek’s well-known arguments against the possibility

of an adequate scientistic methodology – arguments intimately connected to his criticisms of

both socialism and Keynesian-style aggregate demand management – are closely related to the

possibility of ecologically-rational outcomes within the context of scientific discovery and,

therefore, to the need for a degree of liberty within this context. Scientism, for Hayek, was the

view that the social sciences could be (and could only be) deliberately reconstructed in order to

generate predictions of a degree of specificity and accuracy necessary to facilitate effective

socialism and Keynesian demand management. However, Hayek argued, scientistic

methodology, not unlike these latter economic-political programs, assumes the sufficiency of

constructivism and the impossibility of ecologically-rational social outcomes. But, according to

Hayek, the defender of scientism is wrong on both counts. The predictive capacity of the social

sciences is constrained by the limits of human cognition with respect to social phenomena. The

scientistic program cannot overcome this hurdle. Moreover, if ecologically-rational scientific

outcomes are possible, the scientistic prejudice that the conscious application of reason is

necessary for the realization of the relevant predictions is misplaced: it is possible that a

spontaneously-evolved social science may discover means to predictions of the required degree

of specificity without resort to the conscious application of reason. Thus, to deny both tenets of


6

scientism is to recognize both the possibility of ecologically-rational scientific outcomes within

the context of discovery and the need for freedom in such cases; it is, in short, to be a

methodological liberal. Hayek’s implicit arguments for methodological liberalism, which the

present paper aims to exposit, are thus intimately connected with his explicit arguments for

liberalism in economics and politics.

2. Complex Orders and Their Investigation

2.1. Complex Orders and Complex Social Orders

According to Hayek ([1964b] 2014, 262), complex phenomena are those that consist of a large

number of elements the activities of which bear such relations both to each other and to the

external environment as to manifest an emergent order that possesses “certain general or abstract

features which will recur independently of the particular values of the individual data, so long as

the general structure…is preserved.”5 Hayek ([1967a] 2014, 278–279) emphasized the

importance of distinguishing between the systems of rules followed by the elements of a

5 However, it should be noted that Hayek’s definition of complex phenomena, perhaps strangely, runs in epistemological rather than ontological terms. That is, complex phenomena, according to Hayek ([1964b] 2014, 260–261), are those the explanation of which requires a model consisting of a large number of variables.

[T]here seems to exist a fairly easy and abstract way to measure the degree of complexity of different kinds of abstract patterns. The minimum number of elements of which an instance of the pattern must consist in order to exhibit all the characteristic attributes of the class of patterns in question appears to provide an unambiguous criterion…[W]hen we consider the question from the angle of the minimum number of distinct variables a formula or model must possess in order to reproduce the characteristic patterns of structures of different fields (or to exhibit the general laws which these structures obey), the increasing complexity as we proceed from the inanimate to the (‘more highly organized’) animate and social phenomena becomes fairly obvious.

Stated another way, our knowledge of complex orders is limited by definition. Systems of complex phenomena are just those with respect to which our knowledge is limited and the degree of complexity of some phenomena is just a reflection of the extent of this limitation in the particular case.


7

particular group and the orders (or “patterns”) of activity that may emerge at the group level

from the observance of such systems of rules. Given Hayek’s ([1967a] 2014, 278) definition of a

“rule” as a “statement by which a regularity of the conduct of individuals can be described,

irrespective of whether such a rule is ‘known’ to the individuals in any other sense than that they

normally act in accordance with it” [italics added], the elements of an order may be of any

degree of consciousness. The activities of non-conscious “galaxies, solar systems…and social

orders”6 as well as those of conscious organisms7 (Hayek [1967a] 2014, 286) cannot be fully

reduced to the properties of their elements, but rather emerge from the regular interactions both

between various proper subsets of their parts and between the (proper and improper) subsets of

their parts and the environment. Social orders are those unique orders the elements of which are

conscious organisms.

The rules of conduct the observance of which will, under appropriate circumstances, give

rise to an emergent social order may be transmitted either genetically or culturally (Hayek

[1967a] 2014, 278). Among order-generating, culturally-transmitted, systems of rules Hayek

distinguished between those that are deliberately designed and imposed, and those that evolve in

an undesigned process over a (typically long) period of real historical time. Hayek denoted

designed orders produced by imposed systems of rules as “organizations” (Hayek [1967a] 2014,

278n) and undesigned orders that emerge from evolved systems of rules as “spontaneous”

6 “[S]patial pattern[s] such as will occur in the marching, defence, or hunting of a group of animals or men. The arrow formation of migrating wild geese, the defensive ring of the buffaloes, or the manner in which lionesses drive the prey towards the male for the kill, are simple instances in which it is presumably not an awareness of the overall pattern by the individual but some rules of how to respond to the immediate environment which co-ordinate the actions of the several individuals. More instructive are the abstract and more complex orders based on a division of labor which we find in such insect societies as those of bees, ants, and termites.” (Hayek [1967a] 2014, 281) 7 The mind of a conscious organism (i.e., its “sensory order”) is, according to Hayek (1952, 16), “a particular order of a set of [neural] events taking place in some organism and in some manner related to but not identical with, the physical order of events in the environment.”


8

(Hayek [1967a] 2014, 286).8 The social orders that we observe in practice are typically products

of mixtures of genetically-transmitted and culturally-transmitted – some imposed and some

evolved – systems of rules.9

It is not the case that the observance of just any system of rules will give rise to an

emergent social order. Whether and to what extent an order emerges from conduct guided by a

system of rules depends on the degree to which this system is fitted to the environmental

circumstances in which the members of the relevant group act (Hayek [1967a] 2014, 280). A

system that includes the rule “murder all newborns” is probably not conducive to the emergence

of a social order under any circumstances, while some systems of rules may promote the

emergence of an order, say, under peaceful conditions but not under conditions of war, or in a

free society but not in a totalitarian dictatorship, or in some technological circumstances and not

in others, or in good weather but not in bad, etc. Social disorder is thus a consequence of either

an absence of rules that guide the activities of individuals or the inappropriateness of prevailing

systems of rules of conduct in the particular environment in which members of the relevant

group act (Hayek [1967a] 2014, 279–280).

Moreover, social order is a matter of degree, i.e., a question of the extent to which the

system of rules governing the activities of the members of a group is adapted to prevailing

circumstances, and will change over time with changes in the environment (Hayek [1968] 2014,

308–309). The survival of a group of organisms depends on the orderliness of its society and,

8 The “chiefly negative (or prohibitory) rules of conduct which make possible the formation of social order are of three kinds…(1) rules that are merely observed in fact but have never been stated in words[;]…(2) rules that, though they have been stated in words, still merely express approximately what has long before been generally observed in action; and (3) rules that have been deliberately introduced and therefore necessarily exist as words set out in sentences” (Hayek [1970] 2014, 343–344). 9 “[T]he degree to which outcomes are unintended is a continuum” (Schmidtz 2012).


9

thus, requires that individual members adhere to rules of conduct that are fitted to the prevailing

environmental circumstances. Given that these latter conditions are in constant (albeit more or

less gradual) flux, the persistence of a social order (and, thus, of its members) requires continual

adaptation of the system of rules that guide the activities of its members (Hayek [1967a] 2014,

280, 282).

A change (“mutation”) in the rules according to which the members of a group act is

progressive if and only if it contributes to an increase in the group’s numbers across time (Hayek

[1970] 2014, 345; also [1967a] 2014, 288). This might mean that individuals join the group who

were outside of it prior to the mutation or that the mutation leads to more members being born

into the group (and surviving) than pass out of the group via death or attrition.

Although Hayek was not specific on this point, the definition of a progressive change in

the rules that give rise to a social order can be (and perhaps should be) relativized to the growth

of the population of which the group in question may merely be a proper subset. We might

hesitate to describe some rule change as progressive if, despite the fact that it increases the

numbers of its subscribers, the numbers of the broader population increase more than those of the

group over the relevant time span. We can, then, in like fashion, define a static rule change as

one that merely maintains the group’s numbers in a constant ratio with those of the broader

population and a regressive mutation as one in which the proportion of group members shrinks

relative to the population.


10

The behavior of an individual organism is never fully determined by the (social) rules it

follows, but always in conjunction with its own internal impulses and external stimuli.10

Individual members of the same group will thus at any given moment follow different rules as

these latter considerations dictate or “because different rules apply to different individuals

according to age, sex, status, or some particular state in which each individual finds itself at the

moment…[R]ules of conduct will always act only as a restraint on actions induced by other

causes” (Hayek [1967a] 2014, 280). Whatever the cause of some particular action of an

individual member of a group, that it is part of a system of actions that constitute an order will be

indicated by the cessation of the motivating force of the initial cause of the activity (Hayek

[1967a] 2014, 280). In other words, whatever craving or discomfort might lead an individual to

act in particular circumstances, if she lives in an orderly society and behaves in accordance with

the system of rules of conduct from which that social order emerges, her actions and those “of

the different individuals will be so coordinated, or mutually adjusted to each other, that the result

of their actions will remove” the motivating uneasiness (Hayek [1967a] 2014, 280). Whatever it

is that stimulates my desire for prodigious quantities of sweet pastries and whatever reasons

inspire the proprietor of my local patisserie to sell sugary bits of baked dough, the fact that we

live in a social order that emerges from the observance of certain rules (which neither I nor my

local pâtissier need know explicitly or be able to state explicitly), facilitates the fulfilment of our

respective yearnings.

10 The parenthetical “social” is necessary because, if Hayek’s (1952) cognitive psychology is sound, then an organism’s internal impulses emerge from the rules which, together with external stimuli, determine its sensory order. An organism’s behavior would thus be the results of the interaction of external stimuli and systems of rules on multiple levels.


11

2.2. The Methodology of Sciences of Complex Phenomena

The scientific investigation of complex orders raises special difficulties. The aim of the relevant

sciences is the reconstruction of such orders. The social sciences, for example, aim to show how

social orders emerge from the “interplay of the rules of conduct of” individual members of the

relevant group “with the actions of other individuals and the external circumstances” (Hayek

[1967a] 2014, 283). However, our knowledge of complex emergent orders – or, more exactly,

our ability to reconstruct them on the basis of knowledge of the properties of their elements, the

interrelations that obtain between the various proper subsets of these elements, and the external

connections of these (proper and improper) subsets, not to mention our knowledge of the

relevant data, i.e., our knowledge of the values the theoretical variables assume at any particular

time and place – is necessarily limited compared to what we can know about systems explicable

in terms of models consisting of relatively few variables which, for the purposes of analysis, can

be treated as closed with regard to the environment (Hayek [1964b] 2014, 261–262). It is the

defining methodological character of such simple phenomena that “it will generally be possible

to specify all those aspects of the phenomenon in which we are interested with any degree of

precision which we may need for our purposes” (Hayek [1955] 2014, 200), but this is not the

case with respect to complex structures that emerge from the interactions of a large number of

variables with their environment. Here it is typically the case that either our theory of these

interactions is partial and vague, i.e., “more remote from reality—requiring much more

additional knowledge before [it] can be applied to particular instances” (Hayek [1956] 1967,


12

123), or our ability to acquire the data necessary for a more complete explanation of relevant

events is deficient (Hayek [1945] 2014, 95–100).

Stated differently, for Hayek, knowledge is of two kinds: theoretical knowledge of

general rules (and their internal and external interactions) and empirical knowledge of “particular

circumstances of time and place” (Hayek [1945] 2014, 95–96), and because of the multiplicity

and structural complexity of the relations obtaining between the former, and because the latter

are both equally abundant / complex and often exist “only as the dispersed bits of incomplete and

frequently contradictory knowledge which all the separate individuals possess” (Hayek [1945]

2014, 93) – knowledge which will often be held only tacitly by those who act upon it (Hayek

[1963a] 2014, 233–234)11 – the scientist of complex phenomena cannot construct explanatory

models that approach the comprehensiveness of explanations of systems of simpler phenomena.

Explanations and predictions are always a matter of “degree,” i.e., “they will always state

only some and never all the properties of any particular phenomenon to which they refer[,] [and]

each property stated will be expressed not as a unique value or magnitude but as a range,

however narrow, within which the property will fall” (Hayek [1955] 2014, 201). That is,

explanations and predictions prohibit the occurrence of events that fall outside a particular range,

and are undermined, if not falsified, by the appearance of events they forbid. This is true even in

sciences that deal with comparatively simple phenomena (Hayek [1955] 2014, 201) but, because

of our cognitive limitations, our efforts to reconstruct complex orders via theoretical models

necessarily represent less of the relevant phenomena than our models of simpler phenomena

11 Hayek emphasized the crucial role in the emergence of social orders of what Polanyi (1966) called “tacit” knowledge and what Ryle (1946) called “knowledge how” as opposed to “knowledge that”—that is, in Hayek’s (1952, 19) words, knowledge with regard to which we may not be “explicitly aware,” but which we “merely manifest…in the discriminations which we perform.”


13

(Hayek [1955] 2014, 209–212),12 and thus, the best explanations and predictions that we can

manage with respect to the former are necessarily of lesser “degree” than our best explanations

of simpler phenomena, i.e., the former prohibit a smaller range of phenomena than the latter.13

Moreover, any attempt to model complex phenomena as if they were simple is likely to

underwhelm our explanatory ambitions (Hayek [1964b] 2014, 263).14 Hayek’s methodological

perspective suggests that a possible reason for the failure of a particular explanation is that the

relevant phenomena is modeled as simple despite requiring a complex model, i.e., that the

particular model (or, more precisely, the assumed meta-model—the model of an appropriate

model) is ill-suited to an explanation of the relevant phenomena.

For the social scientist, it may be either that the order itself is more readily accessible to

observation than the rules from which it emerges or vice versa. That markets (Hayek [1964b]

2014, 269–271), natural languages (Hayek [1963a] 2014, 232), and perhaps science itself exhibit

certain regular patterns in the activities of their constituents is probably easier to recognize than

the rules the observance of which give rise to these patterns. For Hayek, then, the business of the

economist and linguist (and, perhaps, the social scientist of science) is the genesis and testing of

hypothesized systems of evolved rules of conduct obedience to which could give rise to the

12 Again, this is an implication of Hayek’s definition of complex phenomena in terms of the complexity of the model required to explain some phenomena, and does not, by itself, imply that there are sciences of complex phenomena. The question whether some (or, indeed, any) discipline qualifies as a science of complex phenomena in Hayek’s sense is an empirical one. Hayek included among the sciences of complex phenomena theoretical psychology (1952), economics, linguistics ([1967a] 2014, 283), geology, evolutionary biology, and the branches of astrophysics that investigate the formation of stars and galaxies ([1967a] 2014, 286); as well as “cybernetics, the theory of automata or machines, general systems theory, and perhaps also communications theory” ([1955] 2014, 211). 13 With respect to complex phenomena, our explanations are of “the principle[s] on which a certain mechanism operates” (Hayek [1964b] 2014, 272) and our predictions are of “patterns” in the relevant events rather than of the specific events themselves (Hayek [1964b] 2014, 263–264). 14 “[A] simple theory of phenomena which are in their nature complex [i.e., which can be adequately explained only on the basis of a complex model]…is probably merely of necessity false—at least without a specified ceteris paribus assumption, after the full statement of which the theory would no longer be simple” (Hayek [1964b] 2014, 263).


14

observed order under the relevant environmental circumstances.15 On the other hand, with

respect to primitive societies, “it is often easier to ascertain the rules of individual conduct than

to trace from them the resulting overall and often highly abstract order. The individuals will

often themselves be able to tell us what they regard as appropriate action in different

circumstances, though they may be able to do this only for particular instances but not to

articulate the rules in accordance with which they act” (Hayek [1967a] 2014, 281–282). The

anthropologist’s task is thus to refine her subjects’ statements regarding proper action in various

circumstances into statements of general rules of conduct and to reconstruct “the overall order

which is produced by actions in accordance with them” so as to discover “the [social]

‘functions’16 which these rules serve” (Hayek [1967a] 2014, 282).

Hayek ([1964a] 2014, [1967b] 2014) associated systems of complex phenomena with a

bifurcation in the nature of philosophical rationalism. In particular, he argued that it was

appropriate to adopt the traditional Cartesian – “constructivist” – attitude toward the human mind

as a source of knowledge only with respect to those phenomena that could be adequately

explained in terms of a deliberately reconstructed model. That is, constructivist rationalism is an

appropriate attitude only with respect to simple phenomena. But, with respect to complex

phenomena, constructivism presupposes more extensive knowledge than the human mind can

possess (Hayek [1964a] 2014, 42–43), and, thus, complex orders can at best be only partially

15 Such a hypothesized system of rules of conduct might be tested “by seeking for yet unobserved consequences which follow from it, and by asking whether all [orders of the kind the system is postulated to explain] which we find can be accounted for” by it (Hayek [1967a] 2014, 287). 16 Social orders “exist because the elements do what is necessary to secure the persistence of that order. The ‘final cause’ or ‘purpose’, the adaptation of the parts to the requirements of the whole, becomes a necessary part of the explanation of why structures of the kind exist…A ‘teleological’ explanation is thus entirely in order so long as it does not imply design by a maker but merely the recognition that the kind of structure would not have perpetuated itself if it did not act in a manner likely to produce certain effects, and that it has evolved through those prevailing at each stage who did” (Hayek [1967a] 2014, 288).


15

reconstructed. Adopting an attitude of Cartesian constructivism with respect to a complex system

– i.e., treating it as amenable to a comprehensive reconstruction – is likely to do more harm than

good (Hayek [1964a] 2014, 42).17 Where the human mind is necessarily limited in the

knowledge it can possess and is thus not able to fully reconstruct some phenomena, the

appropriate attitude is the “anti-rationalism” of David Hume and the Scottish Enlightenment

philosophers, who “pleaded that an effective use of reason required a proper insight into the

limits of the effective use of individual reason” (Hayek [1964a] 2014, 41).18

To sum up, constructivism is limited by the requirement that everything relevant to a full

explanation of some phenomena be known. But, beyond this limit, where rational outcomes

require an ecological element, the attitude of constructivist rationalism is, at best, impotent, and,

at worst, deleterious. Within the range of ecological rationality, outcomes are such that if

everything relevant to a full explanation were known or knowable, the phenomena could be

recognized as their necessary consequence, but which may not be so recognized precisely

because the relevant facts are not all known and, perhaps, are not all knowable.

3. Science as a Complex Social Order

3.1. Properties of a Complex Scientific Order

17 With respect to constructivist philosophy of science McQuade (2010, 45) writes: “the proposal of prescriptive injunctions appeals to those with a constructivist bent, and in this guise they can be positively damaging if imposed (even with the best of intentions) on an adaptive system from the outside[.]” 18 In harmony with the writings of his friend Karl Popper, Hayek described this latter attitude as one of “critical rationalism” ([1964a] 2014, 53). Subsequent authors (in particular, as mentioned above, Vernon Smith (2008)) have described it as an attitude of “ecological” rationalism. The latter locution is adopted here on the grounds that it better expresses what, for Hayek ([1964a] 2014, 43-44), is the crux of the distinction, in particular, that constructivists take society and its institutions to be the products of human reason – which is thus assumed to be unbounded in its capacity for knowledge and control of these institutions – while the attitude of the ecological rationalist takes human reason to be an (at least partially) evolved product of society’s development and thus constrained by this development.


16

Social orders are unique in that their elements – conscious organisms or, more exactly, the minds

of organisms – are, according to Hayek’s (1952) theoretical psychology, themselves complex

orders “whose chance to persist depends on (or at least is improved by) their being part of the

more comprehensive [social] structure” (Hayek [1967a] 2014, 288).

“We have to deal here with integration on at least two different levels, with on the one

hand the more comprehensive order assisting the preservation of ordered structures on the

lower level, and, on the other, the kind of order which on the lower level determines the

regularities of individual conduct assisting the prospect of the survival of the individual

only through its effect on the overall order of the society. This means that the individual

with a particular structure and behaviour owes its existence in this form to a society of a

particular structure, because only within such a society has it been advantageous to

develop some of its peculiar characteristics, while the order of society in turn is a result

of these regularities of conduct which the individuals have developed in society.” (Hayek

[1967a] 2014, 288; italics added)

If it is ever the case that the activities of scientists form (to some extent or other) a social

order – then it is a consequence of individual members of the scientific community conforming

to rules of conduct which, in both their mutual interrelations and external connections with the

environment, suffice for the emergence of such an order. Following Hayek ([1967a] 2014, 279),

if we define the different kinds of elements of a group in terms of the rules that they obey, then


17

we can define scientists as a group of individuals who adhere to a particular system of rules of

conduct in their science-related activities. Likewise, we can define, say, natural scientists (or

social scientists) as a group of individuals who in their relevant activities follow a particular

system of rules of conduct that overlaps the system of rules of conduct from which the broader

scientific order, if it exists at a particular time and place, emerges. This same classification

procedure can be applied at the level of the members of particular disciplines within the natural

and social sciences, and, again, at the level of the members of particular research traditions

within these disciplines and, yet again, to the members of any sub-traditions within these

research traditions. Such a conception implies an image of the relations between systems of rules

of conduct that are observed on various levels of scientific activity that might look something

like this:

Scientists: All follow: A = R1, R2, …, Ra

Natural Scientists: All follow: B = A, RNS1, RNS2, …, RNSb

Physicists: All follow: C = B, RP1, RP2, …, RPc

String Theorists: All follow: D = C, RST1, RST2, …, RSTd

Loop Quantum Gravity Theorists: All follow: E = C, RLQG1, RLQG2, …, RLQGe

Chemists: All follow: F = B, RC1, RC2, …, RCf

Organic Chemists: All follow: G = F, ROC1, ROC2, …ROCg

Inorganic Chemists: All follow: H = F, RIC1, RIC2, …RICh


18

And so on with regard to other natural-scientific disciplines, research traditions, and yet finer-

grained distinctions;

Social Scientists: All follow: I = A, RSS1, RSS2, …, RSSi

Economists: All follow: J = I, RE1, RE2, …, REj

Austrian Economists: All follow: K = J, RAE1, RAE2, …, RAEk

Hayekian Austrians: All follow: L = K, RHA1, RHA2, …RHAl

Misesian Austrians: All follow: M = K, RMA1, RMA2, …RMAm

Keynesians: All follow: N = J, RK1, RK2, …, RKn

Post-Keynesians: All follow: O = N, RPK1, RPK2, …, RPKo

New Keynesians: All follow: P = N, RNK1, RNK2, …, RNKp

Sociologists: All follow: Q = I, RS1, RS2, …, RSs

Marxist Sociologists: All follow: R = Q, RMS1, RMS2, …RMSr

Durkheimian Sociologists: All follow: S = Q, RDS1, RDS2, …, RDSs

And so on with regard to other social-scientific disciplines, research traditions, and yet finer-

grained distinctions.

What this table19 is meant to illustrate is that if there is a scientific order that occupies

some district of a broader social order, then it requires there to be a system of rules (A) that is

19 It should perhaps be emphasized that the table is offered as a heuristic device and not as a collection of postulates. That is, the table shows one way in which the rules (if they exist) that govern various kinds of scientific activity might be related to each other and should not be construed as a commitment to the claim that they in fact both exist and are so related to each other. The table should not be construed so as to commit the author, e.g., to the claim that there is some system of rules which social scientists follow in virtue of which they are scientists just like their compatriots in the natural sciences.


19

followed by some individual members of society, who we call “scientists.” Similarly, the table

indicates that if there is a natural-scientific order within the broader scientific order, then it

requires there to be a system of rules that is followed by some of the scientists, who we call

“natural scientists.” Given that these natural scientists are scientists, then this latter system

encompasses the rules that all scientists follow (i.e., A) and then some. Likewise, the table shows

that if there is a physical-scientific order within the broader natural-scientific order (and, thus,

within the still broader scientific order), then there must be a system of rules that is followed by

some natural scientists, who we call “physicists.” Again, since these physicists are natural

scientists and, therefore, scientists, they share certain rules (i.e., B) with all other natural

scientists and, thus, with all other scientists (i.e., A). And so on.

Of course, the mere existence of some such system of rules, while necessary for the

presence of a social order at the relevant level of scientific activity, does not suffice for the

emergence of a social order at that level; this requires both that the various proper subsets of

such a system be appropriately related to each other and that the various (proper and improper)

subsets of the system be sufficiently adapted to prevailing environmental circumstances. These

latter conditions include those of broader social orders in which the system is embedded as well

as those of the non-social (i.e., physical) environment. That is, the emergence of a scientific

order requires that the relevant system of rules of conduct (A) be adapted not only to physical

circumstances, but also to the system of rules of conduct from which the social order emerges—

the rules of scientific conduct must adapt both to social and non-social circumstances.20

20 “We view science as a complex order comprised of components (individual scientists)…whose behavior is situated and constrained within an institutional structure of routines, conventions, and organized groupings that are themselves ordinarily and substantially emergent phenomena of the order itself” (McQuade and Butos 2005, 346; italics added).


20

Something similar can be said about any orders that might emerge at lower levels of

scientific activity: i.e., among the circumstances to which a system of rules of conduct must be

adapted if it is to give rise to, say, an order in the activities of natural scientists are the rules the

observance of which might give rise to an order in the activities of the broader community of

scientists. Likewise, among the circumstances to which a system of rules of conduct must be

adapted if it is to give rise to an order in the activities of physicists are the rules that give rise to

an order in the activities of the broader community of natural scientists; and so on down the

various levels of scientific activity to the members of particular disciplines, research traditions,

sub-traditions, etc.

This is to say that if the observance of any such system of rules is so internally and

externally well-adapted as to give rise to an emergent social order at the relevant level of

scientific activity, we will be dealing with integration between orders on at least three levels, i.e.,

between the comprehensive social order that facilitates the persistence of organisms, the mental

order of the individual organism that determines its conduct, and an order on at least one

intermediate level—i.e., that of the sub-tradition (or research tradition or scientific discipline,

etc.). This intermediate order, too, will facilitate the persistence of the organism in a particular

form (i.e., qua string theorist or qua physicist, or qua natural scientist, etc.), while, at the same

time, both contributing its part to the emergence of the broader social order and being dependent

upon the rule-guided conduct of certain individuals. The survival of an order at an intermediate

level of scientific activity between that of the individual scientist and that of society requires

both adequate membership from the lower level of individuals and adaptation both to the

physical environment and to any and all systems of rules that exist at higher levels of activity. In


21

short, if the phenomena of scientific activity are complex in Hayek’s sense, then the individual

scientist owes her existence in a particular form to multiple levels of social orders to whose rules

she subscribes, because only within such orders has it proven felicitous to develop her unique

characteristics, while a social order at any particular level persists only because both individuals

act in accordance with the rules from which it emerges and it is sufficiently well-adapted to its

environment, including any other systems of rules of conduct in the vicinity.

The absence of a pattern in the activities at a particular level of the members of a group of

scientists – i.e., disorder at the relevant level – is thus a result either of an absence of a system of

rules of conduct or the presence of an ill-adapted system of rules at the relevant level of activity.

The extent to which the activities at a particular level of the members of a group of scientists are

ordered is a matter of the extent of adaptation to conditions in both the social and more narrowly

physical environment of the system of rules that guides the conduct of individual scientists at

that level.

The definitions of progressive, static, and regressive rule changes discussed in the

previous section are – with certain qualifications – readily applicable to the problem of the

orderliness of scientific activity. A progressive (full stop) change in the rules of conduct that give

rise to an order in the activities of scientists at some level is one that increases the numbers of the

group of scientists conforming to the mutated system of rules. Static and regressive mutations are

those that, respectively, maintain and decrease the numbers of the group of scientists conforming

to the modified system of rules.21 More expressive notions of progressivity, stasis, and

21 We might improve upon these mere “counting” notions of progressivity, etc., if we could somehow assign weights to the numbers that reflect the relative degrees to which different systems of rules of scientific conduct are accepted by individual scientists. However, the results of such an extended analysis will ultimately depend on the mathematical figure selected to represent these cognitive measurements and, given the current state of


22

regressivity follow from considering the consequences of a mutation of a system of rules of

scientific conduct in the light of changes in the social environment. In particular, we might say

that a mutation of the rules of conduct that gives rise to an order in the activities of a group of

scientists is progressive (static or regressive) relative to changes in the memberships of groups of

scientists at either the same or higher levels of activity. For example, the rules of conduct of a

research tradition might change so as to affect the numbers of its members relative to the

numbers of members of either other research traditions within the same discipline, the discipline

itself, natural science (social science), or all of science.

Relativizing the notion of progressivity in this way permits more nuanced judgments

concerning the condition of an order at some level of scientific activity. For example, a research

tradition that is progressing simpliciter may or may not be progressing relative to the changes in

the membership of some other research tradition within the same discipline. That is, a scientific

order at some level of activity may be progressive simpliciter (if its membership is expanding)

but regressive relative to one or more of its rivals (if the memberships of the latter are growing at

a faster rate) or vice versa. Similarly, there may be contexts in which judgments are required

concerning the progressivity of, say, a research tradition relative to the consequences of changes

in the rules of conduct at the higher level of the discipline. That is, a research tradition may be

progressive (static or regressive) relative to the progressivity (stasis or regressivity) of the

discipline to which it belongs. Indeed, yet more complicated judgments are possible given the

conception of science as a complex social order: it may be, for example, that a research tradition

methodological knowledge, there’s little reason to suspect that any one of the myriad mathematical figures that might represent these weights actually expresses them more accurately than some other figure. In other words, though it might be desirable to ultimately extend these mere “counting” concepts, such an extension could be based, in the current state of knowledge, on nothing more profound than a pretence of knowledge.


23

is progressive simpliciter, regressive relative to at least one other tradition in the same discipline,

progressive relative to the discipline itself (because membership in the tradition is increasing at a

rate faster than that of the discipline), regressive relative to natural / social science (because

membership in the tradition is increasing at a rate slower than that of natural / social science),

and progressive relative to science itself (because membership in the tradition is increasing at a

rate faster than that of the most comprehensive scientific order). Thus, the treatment of science as

a complex social order affords intricate and highly contextualized judgments as to the

progressivity of different orders of scientific activity.

The progressivity, stasis, and regressivity of mutations in systems of rules of conduct are

inherently dynamic notions, but we can derive from them synchronic concepts concerning the

condition of an order at a particular time. That is, we can define the dominant order on a

particular level of scientific activity at a given time as the one with the largest membership at

that time. Subordinate orders are then just orders on the same level of activity with smaller

memberships. These synchronic concepts can then be combined with the dynamic notions of

progressivity, stasis, and regressivity, and deployed in yet more complex judgments of the

condition of an order on a particular level of scientific activity. A research tradition, say, might

then be evaluated for the degree of its dominance and the extent of its progressivity relative to

either other research traditions, its discipline, natural science / social science, or science itself.

3.2. A Hayekian Social Science of Science


24

If there is ever an order to some degree or other in the activities of scientists, then there is an

object for a Hayekian social science of science that would aim to determine the extent of such

orderliness and provide explanatory principles of the mechanism by which it emerges from the

interaction between the observation of a system of rules of conduct by a group of individuals and

the prevailing circumstances of the social and physical environment. Beyond this, the Hayekian

social scientist of science might seek an explanation of the extent of disorderliness at some level

of activity in either the absence or environmental incompatibility of a particular system of rules

of conduct. Such a scientist might further aim to distinguish the environmental circumstances

most conducive to the emergence of scientific order in general (i.e., more or less regardless of

the rules observed) from those conditions in which the emergence of an order would require the

observance of a particular system of specific rules of conduct. Similarly, the Hayekian social

scientist of science might aim to identify the principles of operation of systems of rules of

conduct which, under the environmental conditions then prevailing, would give rise to an

emergent scientific order. In short, a Hayekian social science of science would involve the

development of models of hypothesized environments consisting of physical and social

circumstances, and their marriage with postulated systems of rules of conduct, and the drawing

out of implications therefrom, which would then be tested empirically via comparison with

actual circumstances.

Naturally, given the complexity of the phenomena, the Hayekian scientist of scientific

order would be in an unenviable epistemic position relative to the one she would occupy if

scientific phenomena were simpler. Her ability to reconstruct an observed scientific order would

depend on her limited knowledge of the relevant rules, the interrelations between these rules, and


25

their relations to environmental circumstances. However, the scientific order itself may be more

obvious to the observer than the rules from which it emerges. These rules may be largely tacit to

the scientific actor and, thus, mostly inaccessible to the Hayekian social scientist of science. The

capacity of the latter to explain particular events within or manifestations of this order would

depend on her ability to know the relevant data, i.e., the circumstances to which these rules are

applied at particular places and times. The limits of human knowledge with respect to complex

scientific phenomena require the scientist of scientific order to aim lower in her explanatory

ambitions; thus, her limitation to explanations of the principle of the mechanism by which

scientific orders might emerge in appropriate circumstances. Like all explanations, these will

prohibit the occurrence of events that fall outside a particular range of phenomena to which they

refer (and will be undermined, if not falsified, by the appearance of events they forbid). But,

because of our epistemic limitations, our best explanations and predictions with respect to

scientific orders will necessarily be of lesser “degree” than our best explanations of simpler

phenomena.

Given the limitations of the social scientist of science with respect to the relevant

phenomena, a constructivist approach to complex scientific orders, i.e., an attempt to model the

complex phenomena of a scientific order as if they were simple, and, thus, fully and explicitly

explainable, will eventually and inevitably come up short as an account of rational scientific

belief. All of this means that, if scientific rationality includes an ecological component, the

attitude of the constructivist rationalist is inappropriate in methodological investigations. More

exactly, since constructivist and ecological rationality operate together, pure constructivism,


26

unleavened by an appreciation for the possibility of ecological scientific rationality, will

ultimately lead the methodologist astray.

4. Twentieth Century Explanations of Science from the Perspective of a Hayekian

Meta-Methodology

The possibility that some scientific beliefs are rational only in an ecological sense contributes to

an error theory of the shortcomings of previous attempts to explain scientific rationality. That

science is more rational than other methods was typically treated as a datum by earlier

empirically-inclined philosophers of science. However, these same philosophers typically

assumed that the rules of scientific conduct that account for this superior rationality could all be

consciously discovered and discursively stated. In other words, it was generally assumed that

scientific beliefs are rational only in a constructivist sense.22

Constructivism was part and parcel of the philosophies of certain members of the Vienna

Circle—the authors of the Wissenschaftliche Weltauffassung pamphlet, in particular, come to

mind.23, 24 One would be hard pressed to find a more representative statement of constructivist

22 The goal of the offending constructivist philosophies of science was to “articulate an objective procedure or set of criteria by which such candidate knowledge [propositions generated by individual scientists] could be justified as being acceptable…without regard to the context in which they were produced” (McQuade 2010, 26). 23 Written in 1929, the “pamphlet is the product of teamwork; [Otto] Neurath did the writing, [Hans] Hahn and [Rudolf] Carnap edited the text with him; other members of the Circle were asked for their comments and contributions” (Carnap, Hahn, and Neurath 1929). However, not all members of the Circle were on board with the “manifesto,” which “alienated [Kurt] Gödel…to the point that he came to the meetings less and less frequently” (Menger 1994, 210). Indeed, Moritz Schlick himself, “to whom the manifesto was dedicated, was less than satisfied with the result. This was first of all because he was not taken by the conception of the circle as a ‘movement’ of any sort, favouring a more modest and more narrowly scientific approach…But it was also because he was distressed by the political tone of the piece, and more specifically by those portions which suggested some sort of alignment of logical positivism with socialism and with the movement for workers’ education in Vienna at the time” (Smith 1994, 9–10).


27

rationalism, as Hayek and others have conceived it, than the statement that “[t]he Vienna Circle

believes that…it fulfils a demand of the day: we have to fashion intellectual tools for everyday

life, for the daily life of the scholar but also for the daily life of all those who in some way join in

working at the conscious re-shaping of life. The vitality that shows itself in the efforts for a

rational transformation of the social and economic order, permeates the movement for a

scientific world-conception too” (Carnap, Hahn, and Neurath 1929); or, again, the notion that

“[t]he scientific world-conception knows no unsolvable riddle” (Carnap, Hahn, and Neurath

1929; italics in the original); or, finally, that “[w]e witness the spirit of the scientific world-

conception penetrating in growing measure the forms of personal and public life, in education,

upbringing, architecture, and the shaping of economic and social life according to rational

principles. The scientific world-conception serves life, and life receives it” (Carnap, Hahn, and

Neurath 1929; italics in the original).

The slogan that scientific justification means verification via logical analysis –

“reduction to the simplest statements about the empirically given” (Carnap, Hahn, and Neurath

1929) – or, for that matter, any appeal to a “naïve” falsificationism (Lakatos 1968–1969, 152– Relations between members of the Austrian School of economics and the Vienna Circle of Logical Positivism were, to say the least, complicated. On the one hand, the mathematician Karl Menger (son of Carl Menger, founder of the Austrian School) and the Husserlian philosopher Felix Kaufmann participated in, and maintained close, in some cases, lifelong, friendships with members of both kreise. On the other hand, there was no love lost between Ludwig von Mises, the arch-liberal political economist around whom much of post-WWI Viennese economics centered, and Otto Neurath, the arch-Marxist “social science expert” of the Vienna Circle. Mises and Neurath had been co-participants in Eugen Böhm-Bawerk’s famous economics seminar before the war, as well as disputants in the German-Language Socialist Calculation Debate after the war (for more on the relationship between Mises and Neurath, see Chapter Five of Caldwell 2004). Indeed, Barry Smith (1990, 220) speculates that Hayek himself may have “seriously considered joining the Vienna circle,” but was dissuaded (one might imagine, under the influence of his mentor, Mises) by Neurath’s “naïve, not to say absurd, economic views.” 24 For a discussion of several further representatives of constructivism in the sciences and philosophy – namely, the American economist Wesley Clair Mitchell, the Canadian scientist G. Brock Chisholm, The German-American philosopher of science Hans Reichenbach, and Hayek’s friend and great rival, the British economist John Maynard Keynes – see footnotes 4 and 23 of Hayek ([1970] 2014). Hayek would have undoubtedly agreed with Weimer’s (1980, 178-179) assessment of Descartes, Rousseau, Comte, Marx, Lenin, Freud, Bertrand Russell, John B. Watson, and B.F. Skinner as devout constructivists.


28

162) that takes a single countervailing observation as sufficient to both conclusively falsify a

universal statement and warrant its rejection – implies an assumption of constructivist

rationality.25 That is, these views imply that scientific rationality requires nothing more than the

mechanical application of some known (or, at least, knowable) rule to explanations of a

particular body of evidence. Indeed, implicit in such accounts is the notion that a (the) system of

rational scientific beliefs could (at least in principle and perhaps in practice) be reproduced in a

laboratory environment.26 All that would be required would be to enforce adherence to the

relevant rule(s) among human test subjects who had been provided with certain bits of evidence

and various competing explanations.

Of course, it was widely acknowledged by the middle of the third quarter of the last

century, largely as a result of the work of W.V.O. Quine ([1951] 1961) and Thomas S. Kuhn

(1962), that actual scientific practice is not so simple. Quine (as Duhem ([1914] 1954) had

previously done) argued that, strictly speaking, no statement is falsifiable in isolation. Thus, the

rule to accept only falsifiable, but not-yet-falsified, statements could be neither a necessary nor a

sufficient criterion of rational scientific belief. Our imaginary test subjects would sooner or later

believe irrationally were they to unwaveringly apply only naïve falsificationism: they would

reject hypotheses which, in the actual course of the history of science, have for one (presumably

good) reason or another been retained in the face of apparent falsification, or they would retain

hypotheses that have in fact been rejected for reasons other than falsification. For his part, Kuhn

25 Weimer (1980, 166) argues – convincingly, to my mind – that even more “sophisticated” versions of Popper’s critical rationalism are constructivist: “[t]he Popperian notion of critical rationalism holds that conduct is rational to the extent it follows clearly specified general rules that can be known to and acted upon by practitioners.” 26 The goal of the defenders of the so-called “Received View” (Suppe 1977) in philosophy of science was “to reduce [scientific] success to a set of logical, even algorithmic rules for grinding out truth claims, sometimes known as Scientific Method” (Leonard 2002, 142).


29

emphasized the social nature of scientific practice: scientific beliefs, such as they are, cannot

merely fall out of the mechanical application of logic to empirical evidence, but result from

processes that necessarily involve psychological and social considerations. There is no

mechanical decision algorithm that might be provided to our imaginary test subjects that would

either necessitate or suffice for their convergence upon a system of rational scientific beliefs.

Yet, for all the credit due to Quine and Kuhn27 for deflating this crude constructivism,

certain of their followers responded by committing what, from the perspective of a Hayekian

meta-methodology, can only be counted as the same error as the one made by the constructivist

empiricists. Rather than recognizing the possibility of rational, but not-fully-effable, scientific

beliefs, some in the sociology of scientific knowledge (SSK) tradition (especially Latour and

Woolgar ([1979] 1986) and, perhaps to a lesser extent, Barnes and Bloor (1982)) simply rejected

the relevant datum: science, they argued, despite all appearances, is not in fact more rational than

other methods of belief fixation. Notice, however, that this leap from the inadequacy of the

aforementioned empiricist philosophies of science to irrationalism is merely constructivism

redux: either the rules of theory choice that constitute science’s superior rationality are fully

describable, or there are no such rules and this apparent rationality is an illusion; science is little

better than witchcraft, religious dogma, or an appeal to political authority as an epistemic

method. The third possibility that the scientific method is in fact epistemically superior to these

alternatives (or, more exactly, that systems of rules of scientific conduct contribute to the

emergence of order under appropriate environmental conditions), but that this epistemic

superiority is at best only partially amenable to deliberate reconstruction, was never seriously

27 And other, similarly-minded thinkers such as Stephen Toulmin (1953, 1961), Michael Polanyi (1958, 1966), Norwood Russell Hanson (1958), and Paul Feyerabend (1965))


30

considered by those writers who leapt from arguments like those of Quine and Kuhn to the

conclusion of epistemic relativism.28 For constructivists – of either an empiricist or a relativistic

bent – it is all or nothing when it comes to science’s rationality.

A Hayekian meta-methodology that conceives of the justification of scientific beliefs as

complex phenomena in the relevant sense implies some minimal criteria for explanations of

scientific rationality. In the first place, against constructivist empiricist philosophers, a Hayekian

meta-methodology implies that science’s superior rationality cannot be taken for granted. To

accept the possibility of ecologically-rational scientific beliefs is to recognize that scientific

beliefs can be more or less rational—i.e., it is to recognize both that scientists can adhere more or

less to an otherwise order-generating system of rules and that such a system can be more or less

well-adapted to prevailing conditions; and, therefore, that various contexts of justification can be

more or less ordered. The matter of science’s epistemic superiority – its degree of orderliness

vis-à-vis other methods of belief fixation – is an empirical question that will likely be answered

differently for unique research traditions, disciplines, etc. at various places and times.

Moreover, the ecological perspective of a Hayekian meta-methodology does not assume

that, if and when adherence to some system of rules manifests a degree of order in the beliefs of

scientists, the relevant rules will all be accessible to cognition. If the phenomena of scientific-

belief justification are complex, then the rules the observance of which give rise to the

emergence of a scientific order cannot all be explicitly stated.

Finally, if the phenomena of scientific practice are indeed complex in Hayek’s sense,

then scientific orders emerge from the interaction of rules that concern scientists’ confrontations

28 Related to this, see Laudan (1990)


31

with both the physical environment and society. That is, an explanation of science that runs

exclusively in terms of either empirical or social considerations but not both is necessarily

deficient on a Hayekian meta-methodology. The trick is to show how scientific orders emerge

from the interrelations between rules of scientific conduct, confrontations between theories and

empirical evidence, and scientists’ own social interactions.

In short, a Hayekian meta-methodology predicts the failure of any methodology that aims

to reconstruct scientific rationality on the basis of either a simple or static set of rules, and the

failure of any explanation of science that denies or ignores either the empirical or social aspects

of the process from which scientific orders emerge. Constructivist empiricist philosophers of

science erred in adopting a meta-model predicated on the comprehensive intelligibility of the

rules that constitute science’s epistemic superiority, and which was, for this reason, unable to

adequately encompass the complexity of the context of justification. But, constructivist

relativists, made the same blunder in leaping from the failure of the empiricist constructivist’s

meta-model to the conclusion that there are no rules of scientific justification, and that science is

merely a matter of psychological and social considerations.29

5. Liberalism in Politics, Economics, and Methodology as Consequences of Complexity

It follows from the foregoing considerations that, where a scientific order emerges despite

limited knowledge of the circumstances that lead to its emergence, i.e., where scientific

rationality is to some degree ecological, coercion can be neither necessary nor sufficient as a

29 McQuade (2010, 41–48) discusses various postpositivist explanations of science which, in his view, better conform to a Hayekian conception of science as an “adaptive classifying system.”


32

scientific method. That coercion is unnecessary under such circumstances is an implication of the

very possibility of the spontaneous emergence of a scientific order. That coercion cannot suffice

under such conditions follows from the absence of the knowledge required to make it effective,

i.e., where the knowledge required to deliberately realize a scientific order is not available, no

amount of force – however brutal the coercive measures employed – will suffice for the

manifestation of an order.

In his political philosophy, Hayek ([1960] 2011, 57) defined liberty as “that

condition…in which coercion of some by others is reduced as much as is possible in society.”

That is, individuals are free to the extent to which their actions are uncoerced by others. Thus,

there is a clear connection between Hayek’s political philosophy and the implications for

scientific order of his writings on complex phenomena.30 To say that scientific rationality cannot

be coerced where it is ecological is to say that the realization of scientific order under such

circumstances requires a degree of methodological liberty.

Moreover, the concept of emergent scientific orders contributes to an explanation of the

context of scientific discovery31 that accords well with Hayek’s argument(s) against the

30 Also see Hayek ([1962] 2014): “[w]e have…learnt that there exist political systems which make very difficult even such intellectual honesty as is a basic condition for all genuine science. It is certainly possible to preserve intellectual honesty in the most difficult conditions. But we are not all heroes, and if we value science[,] we must also advocate a social order which does not make such intellectual honesty too difficult. There seems to me to exist in this respect a close connection between the ideals of science and the ideals of personal liberty.” 31 On the similarities between market competition and scientific method as “discovery procedures,” see Hayek ([1968] 2014, 305). Like competitive methods, “[t]he advantages of accepted scientific procedures can never be proved scientifically, but only demonstrated by the common experience that, on the whole, they are better adapted to delivering the goods than alternative approaches.” The analogy is drawn yet closer, if one accepts Hayek’s ([1946] 2014, 115) argument that “[c]ompetition is essentially a process of the formation of opinion: by spreading information, it creates…unity and coherence…It creates the views people have about what is best and cheapest, and it is because of it that people know at least as much about possibilities and opportunities as they in fact do. It is thus a process which involves a continuous change in the data and whose significance must therefore be completely missed by any theory which treats these data as constant.”


33

methodological variant of constructivist rationalism – a position he dubbed “scientism”32 –

arguments that figure centrally, if not always explicitly, in his well-known criticisms of both

socialism and Keynesian-style aggregate demand management. Indeed, Hayek is probably best

known as one of the twentieth century's most vehement critics of these programs. What is

perhaps less appreciated is that Hayek’s criticisms and, thus, his related defenses of liberalism in

politics and economics, are largely epistemological / methodological in nature.

In the wake of the Great Depression, during the interwar years and after, a wide-ranging

consensus arose to the effect that government planning of the economy was crucial for the future

maintenance and flourishing of western civilization. From the dogmatic Marxism of Neurath, the

“social science expert” of the Vienna Circle, and the gradualist socialism of Fabians like Beatrice

and Sidney Webb, Hayek’s employers at the London School of Economics, to the “Tory

Socialists,”33 who acquiesced to the postwar consensus for the British welfare state, there was

widespread agreement across the political spectrum in favor of some form of central planning.

However, it was also generally acknowledged that the social sciences, as then constituted, were

inadequate to the onerous requirements of successful centralized economic planning. Social

control of the sort required of such planning could only be predicated on highly specified and

reliably accurate predictions of social phenomena. The predictive and explanatory deficiencies of

the social sciences in these respects were too obvious to ignore. The issue for advocates of

economic planning then became the means by which social-scientific predictions might be made

to meet the required standards of specificity and accuracy. Reform of the social sciences, along

32 Although he is often credited with originating the term, according to his reminiscences ([1967] 1992, 173-4), Hayek borrowed the phrase from Otto Neurath, who occasionally used “scientism” to describe his own view. 33 See the proposed chapter outline of an early version of Hayek’s unfinished Abuse and Decline of Reason project reproduced in Caldwell (2010, 5)


34

the lines of the successful techniques of the “hard” sciences, was suggested as a means to the end

of precisified predictions of social phenomena, itself a means to the further end of effective

economic administration of either a socialist or Keynesian variety.

Thus, the scientism that Hayek argued against was the view that the social sciences could

be, and could only be, deliberately reconstructed to deliver predictions of the specificity and

accuracy required of these economic-political programs. Hayek’s argument(s) against scientism

aimed to show that importation of the methods of the physical sciences is no way to improve

predictions of social phenomena and, moreover, that there is no known means for the deliberate

realization of the relevant goal.34 The discovery of a system of rules of scientific conduct capable

of predicting social phenomena with the degree of specificity required for the relevant economic-

political programs – assuming some such system is possible – cannot be deliberately brought 34 Hayek’s understanding of scientism and, thus, the reasons he offered for its inadequacy, changed over time. In fact, he offered two distinct explanations for the inefficacy of attempts to import the methods of the physical sciences into the social disciplines. According to Hayek’s early methodological writings (see especially “Scientism and the Study of Society” [1952] 2010), there is a meaningful difference between social and physical phenomena that makes the methods appropriate to explaining the latter inappropriate for explanations of the former. However, Hayek eventually came to the view, perhaps under the influence of Karl Popper, that this sharp distinction between the social and physical sciences was untenable. According to Hayek's later methodological writings (see especially [1955] 2014 and [1964b] 2014), there is only one scientific method, and it looks more like the causal-genetic approach he had explicated in the "Scientism" essay (see Caldwell 2014, 14) as the spontaneously evolved method of the social sciences than the constructivist suggestions of the defenders of scientism. To the extent this change meant that an appeal to the methods of scientism for the social sciences could not be justified on the grounds of their unique successes in the physical sciences, Hayek’s early argument against scientism was buttressed; but, by the same token, since this change in his conception of the sciences also meant that there’s nothing unique about the social sciences, Hayek’s argument against scientism required some considerable modification. If the techniques of the social disciplines are the same as those of the physical disciplines, why are the predictions of the former so much less satisfactory and practically useful than those of the latter?

Hayek’s answer, foreshadowed in places in “Scientism and the Study of Society,” and emphasized in his later methodological writings, was that the predictive limitations of certain disciplines, especially the social sciences, are to be attributed not to fundamentally different or deficient methods, but to the comparatively higher degree of complexity of the phenomena they investigate. The physical sciences are successful compared to the social disciplines not because their methods are, in some sense, “better” or more worthy of imitation, but because they deal with phenomena that can be explained on the basis of models that are comparatively simple: where the scientific method is applied to phenomena that can be explained with a few variables relatable in terms of experimentally-discoverable laws, it is possible to predict with the degree of specificity and accuracy that we find in the paradigmatic physical sciences. However, where the same method is applied to complex phenomena, it is only possible to explain the “principle” from which the phenomena emerge and predict “patterns” in the phenomena. For more on Hayek’s arguments against scientism, see Caldwell (2010, 35–38).


35

about, but requires an element of spontaneity. That is, Hayek’s conclusion against scientism

constitutes an ignoramus, but not an ignorabimus. From the conclusion that we don’t know how

to deliberately realize the goal of scientism with respect to social-scientific predictions, it does

not follow that this goal will remain forever unrealized. We might, as it were, learn or develop

new methods in the course of other pursuits – or circumstances outside our epistemic purview

might otherwise be sufficiently fortuitous – that the goal is eventually realized despite our

present ignorance. The key point is that, if it is to be realized, it will come about as a result of a

process that we cannot foresee or understand a priori: we cannot plan our way to the required

predictions.35 Thus, Hayek denied both tenets of scientistic methodology: not all scientific goals

can be brought about deliberately and some scientific goals, if they can be realized at all, require

a non-deliberative component.

This is to say that Hayek’s argument is directed against scientism as a means to the

realization of social-scientific predictions of the required degree of specificity and not against

scientism as an end in itself, or as a means toward some other end. The conclusion is not that we

don’t know how to effect scientism in the social sciences – i.e., the conclusion is not that we

don’t know how to act like natural scientists with respect to social phenomena – but that such

pretense is no means to the end of improving social-scientific predictions. The social sciences

can try scientism, and may eventually discover other ends for which it is an appropriate means,

but we should not expect these ends to include that of improving social-scientific predictions to

the extent required of effective political administration of the economy. This latter point implies

35 See Hayek ([1963b] 2014), “[p]erhaps someday we shall get something like a general science of man which will provide us with both the required conceptual framework and the appropriate technique which then as specialists in economics or any other social science can employ as mechanically and unthinkingly as the natural scientists can follow the scientific method he has been brought up to use.”


36

a further possibility, namely, that both the means and ends of science can be rational in an

ecological sense. New methods might develop that allow existing, but previously unachievable,

goals to be realized, or new scientific ends might develop that are realizable with extant methods

that were, however, inadequate for any previous goal.

Again, it follows from these considerations that, where ecological processes are

operative, successful science – the aptness of scientific methods for scientific objectives – cannot

be realized via coercion. And this is so, moreover, for purely epistemic reasons, without regard

to ethical qualms or questions concerning the incentive structure confronting coerced scientists.

If the knowledge required to make effective some plan for the realization of a particular

scientific goal has not been discovered, coercion (intimidation, force, duress, strong-arm tactics,

etc.) can be neither necessary nor sufficient for the discovery of the requisite knowledge. No

amount of violence can make social scientists generate predictions of the degree of specificity

required to facilitate effective political administration of the economy.

And this is to say that some degree of liberty is required in both the pursuit of ends and

the choice of methods in science. This is methodological liberalism. It is essentially identical

with the denial, on the grounds of the possibility of ecologically-rational scientific outcomes, of

the adequacy of constructivist methodology and, thus, of the possibility of coercion as an

effective scientific method.36 Successful science requires that methods be well-adapted to goals

36 It should be noted that the methodological liberal is not a methodological anarchist. The latter shares with the former the denial of both tenets of scientism, but the anarchist, in effect, goes much further in denying the possibilities for a constructivist understanding of some aspects of scientific activity. The methodological liberal, though she is skeptical of the possibility of a comprehensive account of scientific explanation, need not be so nihilistic as the methodological anarchist. The methodological liberal does not subscribe to Feyerabend’s ([1975] 2010, 12) dictum that the only rule of scientific practice is “anything goes.” This is because, as scientific ends are (at least partially spontaneously) realized, i.e., as new scientific orders emerge, strong forces in the form of rules, customs, and traditions will be set up that serve to maintain the existing order. The observance of the rules that inform the activities of the members of some scientific order may be required to secure the persistence of either the


37

(and vice versa). The methodological liberal insists that this fit between scientific ends and the

means for their realization is at least sometimes, to some degree, the process of an evolutionary,

rather than a designed (or designable), process. The methodological liberal denies both aspects

of the scientistic thesis. That is, according to the methodological liberal, it is not the case that

science can always be deliberately changed to serve whatever goals we might adopt. Nor must

science be designed in order to serve our chosen ends. Both the ends and means – the discoveries

and methods – of successful science can, under appropriate circumstances, result from

spontaneous processes. If, at any given time, science (or, more exactly, one of its many

disciplines, research traditions, or sub-traditions) is successful, it is in virtue of a bi-directional fit

between relevant goals and methods that may well be, at least to some degree, unintended. To

insist on the possibility of this result in scientific inquiry – to insist that there are circumstances

in which scientific success cannot be planned and controlled, and, thus, cannot be coerced – is to

be a methodological liberal.

There is another, related route from Hayek's writings on political philosophy to the

conclusion of methodological liberalism. Hayek often argued that the best justification for liberty

in the economic and political domains is the absence of knowledge concerning the circumstances

necessary for the deliberate maintenance, much less the improvement, of society, i.e., that “the

case for individual freedom rests largely on the recognition of the inevitable and universal

order itself or of the scientist within the order. In other words, there will usually exist good (but perhaps only tacitly-known) reasons for scientists within established scientific communities to accept the problems and adopt the methodology associated with their respective community. But, this is not – indeed, cannot – be the case in new areas of inquiry, where the main problem is the discovery of the rules, if any, that will secure an order and its persistence in the relevant environmental circumstances. As Weimer (1980, 163) points out, Feyerabend ([1975] 2010) implicitly accepted “the usual dichotomization between ‘science is rational only if my methodology is accepted’ and ‘if not, science is irrational and therefore anything goes’[.]” Feyerabend, like the constructivist philosophers of science he attacked and the irrationalists he inspired, failed to recognize the possibility of ecologically-rational scientific outcomes. The “anything goes” dictum is, in fact, disguised scientism.


38

ignorance of all of us concerning a great many of the factors on which the achievements of our

ends and welfare depend” (Hayek [1960] 2011, 80). Hayek took it to be empirically obvious that

the deliberate reconstruction of the multitudinous considerations “on which the achievements of

our ends and welfare depend” requires knowledge that extends well beyond the limits of human

cognition and, thus, that such schemes ought to be avoided in the political and economic realms.

That is, according to Hayek, there is no evidence that supports belief in the knowledge and

abilities of those who would aspire to deliberately reconstruct or administer society so as to

increase achievement of human ends and welfare beyond that which manifests in a liberal

society, and so, a preference for liberty should be the default attitude.

An analogous argument for methodological liberty would seem to be defensible on the

basis of the considerations of the present paper. The deliberate reconstruction of the

circumstances that lead to scientific success requires knowledge that extends beyond the limits of

human cognition. If, when, and to what extent the activities of scientists constitute an order, the

rules adherence to which give rise to this order are not all accessible to the methodologist (or

“social scientist of science”). At best, the methodologist can partially reconstruct this order, but –

as its emergence depends on circumstances, e.g., with respect to physical and social conditions in

the relevant environment, which the methodologist cannot possibly know in advance of scientific

inquiry – only after the realization of success. In other words, the methodologist never possesses

the advance knowledge required to ensure the success of some untried scientific inquiry, but can,

at best, pronounce upon what has worked in the past in various circumstances. There is no reason

to believe in the methodologist’s ability to deliberately administer scientific activity so as to

improve the prospects for scientific success beyond that which manifests in an environment in


39

which the activities of scientists are not coerced, and so, a preference for methodological liberty

should be the default attitude. This principle applies not only to the most comprehensive

scientific order, if there is one, but also to each of any orders that might obtain at the level of

disciplines, research traditions, etc.

There is no a priori reason to suspect that science is a one-size-fits-all endeavor and that

a system of rules that has proven successful in one domain can be transferred without cost (or

with benefit) to another in which it has yet to be tried or the results of its application observed.

That is, even after the evidence has been collected concerning the rules that have proven

successful in one domain, the methodologist’s knowledge remains in another sense limited: she

is not in a position to say with a high degree of confidence why such a system has worked where

and when it has. To assert, then, the transferability of one system of rules from a field in which it

has proven successful to either a new environment or a novel field of inquiry is epistemic hubris.

Similarly, if the methodologist is not in a position to declare in advance which existing system of

rules is most appropriate for some proposed scientific inquiry, less so is she able to conjure all of

the possibilities for new systems of rules. Though the methodologist may be able to point to

failed applications of scientific methods to particular objectives in the history of science, there

can be no a priori grounds for the methodologist to prohibit the pursuit of any system of rules

scientists might devise. Again, absent evidence that the methodologist can deliberately improve

upon the workings of spontaneous scientific practice, a preference for liberty should be the

default attitude. Stated another way, in the words of Charles Sanders Peirce, whom Hayek was

fond of quoting: “[u]pon this first, and in one sense this sole, rule of reason, that in order to learn

you must desire to learn, and in so desiring not be satisfied with what you already incline to


40

think, there follows one corollary which itself deserves to be inscribed upon every wall of the

city of philosophy: Do not block the way of inquiry.”37, 38

In closing this section and the main text of the present paper, I’d like to quote an

extended passage from the pen of economist James Buchanan, Hayek’s friend, sometime

colleague,39 and fellow Nobel laureate and defender of classical political liberalism. The passage

appears not in any of Buchanan’s explicitly methodological (or even theoretical) writings, but in

a most unusual place, namely, a letter dated October 17, 1960 addressed to Kermit Gordon (at

the time, a Director in Economic Development and Administration at the Ford Foundation from

which Buchanan and his then-colleagues at the Thomas Jefferson Center for Studies in Political

Economy at the University of Virginia were seeking a grant). Buchanan was at pains to express

the methodological uniqueness of the Jefferson Center vis-à-vis the mainstream of the economics

37 Hayek (1967) adopts this quote, Peirce’s “first rule of reason,” as the epigram of the Politics section of Studies in Philosophy, Politics, and Economics. This would seem to be a further indication of the extent to which, to Hayek’s way of thinking, epistemological / methodological issues were deeply connected with the problems of politics. 38 There is another Hayekian route via which the necessity of the methodologist’s limited knowledge might be established. According to Hayek (1952, 188), no classificatory system can adequately explain itself because any such system must be more complex than that which it classifies: “[a]n apparatus capable of building within itself models of different constellations of elements must be more complex…than any particular constellation of such elements of which it can form a model.” For Hayek (1952, 5), examples of classificatory systems include the minds of organisms (which classify stimuli confronted in an organism’s interaction with the physical environment) and science (which re-classifies the organism’s interactions with the physical environment so as to generate a more reliable model). This means that, if it is to be capable of comprehensively classifying the phenomena of the environment, science must be capable of producing at least as many classifications as there are kinds of phenomena in that environment: a complete scientific explanation of the physical environment requires that science be at least as complex as the environment. But, since methodology classifies science, this means that a comprehensive descriptive methodology must be more complex than science and a fortiori more complex than the physical environment. (And, again, meta-methodology must be yet more complex than methodology, and so on.) But, since the scientist’s knowledge of the complex phenomena of nature and / or society is limited, the methodologist’s knowledge of science is more limited (and the meta-methodologist’s knowledge of methodology is yet more limited, etc.) 39 Regarding Hayek and Buchanan’s personal and professional relationship, see Buchanan (1992). Hayek and Buchanan were together for a semester at the University of Virginia in 1961, briefly at UCLA (Buchanan in the economics department, Hayek a visiting professor in the philosophy department) in 1969–1970, and again as visiting scholars at Stanford’s Hoover Institution in the late-1970s.


41

discipline.40 “There seem to me to be two essential ways of approaching the study of problems of

political, social [and, thus, by the lights of the present essay, scientific], and economic

organization,” Buchanan wrote,

“The first way is that of setting up independently certain criteria or goals for achievement

and to examine existing and potential institutions in the light of their performance or

expected performance in meeting these criteria. This approach, for purposes of exposition

here, may be called the ‘social welfare function’ or ‘social engineering’ approach41…The

second approach is that which deliberately avoids the independent establishment of

criteria for social organization…and instead examines the behavior of private individuals

as they engage in the continuing search for institutional arrangements upon which they

can reach substantial consensus or agreement. It follows from this difference in approach

itself that ‘individual liberty’, in the sense of individual participation in the choices of

appropriate constraints on human action, will tend to assume a necessary, and hence

more prominent, role in the second than in the first.” (Buchanan 1960; italics added)

Later in the same letter, Buchanan acknowledges, as has been stressed in the present essay, that

the two approaches often exist side-by-side. The issue is a matter of emphasis: “Both approaches

to social problems are, of course, appropriate, and in many cases, they tend to coincide and

merge one with the other. The same tools of analysis will be employed in many occasions. The

40 Many thanks to David Levy and Sandra Peart for permission to quote from their (draft) manuscript “‘Almost Wholly Negative’: The Ford Foundation’s Appraisal of the Virginia School,” which includes the Buchanan-Gordon letter in question. 41 Hayek, of course, would have called this constructivism and, when applied to social science, scientism.


42

difference in emphasis is, however, important, even if subtle in any specific context” (Buchanan

1960).

5 Concluding Remarks

The present essay has addressed the implications of the postulate that the activities of scientists

constitute complex phenomena in Hayek’s sense. Three interrelated theses have been defended,

namely, that 1) Hayek’s methodology of sciences of complex phenomena as applied to science

itself implies a meta-methodology, i.e., a set of minimal criteria for explanations of scientific

rationality; 2) this meta-methodology provides an error theory of certain failed explanations of

scientific rationality; and 3) methodological liberalism – the recognition of both the possibility of

ecologically-rational scientific outcomes and, in these cases, of the impossibility of coercion as

an effective scientific method – is an implication of the treatment of science as complex

phenomena and is closely related to his arguments against scientism, which were themselves

fundamental to his arguments against socialism and Keynesian countercyclical policies.

I will close by suggesting a possible avenue for future thought and discussion that

considerations of brevity do not permit me to take up in earnest in the present essay. Namely, the

question whether recognition of the possibility of ecological scientific rationality doesn’t imply

constructivism at the meta-methodological level. To postulate the complexity of the phenomena

of scientific activity is (to some extent or other) to attempt to reconstruct the rules of

methodology and since, by the argument’s own lights, such attempts will always be inadequate if

methodological phenomena are complex, the postulate may well cannibalize itself. On the other


43

hand, perhaps the potentially self-undermining nature of the postulate is more virtuous than

vicious in the sense that it is yet another reminder of our “inevitable ignorance” and of the limits

of knowledge across several related domains, i.e., science, methodology, and meta-methodology.

The postulate essentially asserts that no matter how elaborate our explanations, there will always

remain an unexplained (and, perhaps, unexplainable) residue concerning practice in these fields.

If so, then the possibility of ecological scientific rationality explains (the principle of) its own

limitations. Indeed, in the last analysis, the postulate may be nothing more and nothing less than

a reminder that constructivism is always an inappropriate attitude to adopt toward complex

phenomena.

We should not assume that we possess, or can acquire, all of the knowledge necessary to

explain some phenomena to whatever degree we might like, until experience has shown, with

respect to these phenomena, that we do in fact possess all of the knowledge necessary to explain

the phenomena to the desired extent. The phenomena that satisfy this condition are those we call

“simple” and it is only with respect to these phenomena that constructivist rationalism is an

appropriate attitude. Outside of this realm, where either our knowledge has yet to be tested or has

in fact shown itself inadequate to the desired explanations, the assumption that the deliberate

application of reason is both necessary and sufficient for the realization of our ambitions is an

unfounded article of faith, and the constructivist attitude that ignores the possibility of

ecologically-rational outcomes is bumptious.

References


44

Barnes, Barry and David Bloor. 1982. “Relativism, Rationalism and the Sociology of

Knowledge.” In Rationality and Relativism, eds. M. Hollis and S. Lukes, 21-47.

Cambridge, MA: MIT Press.

Buchanan, James M. 1960. “Letter to Kermit Gordon, October 17, 1960,” William Baroody

Papers, Library of Congress, Box 79, Folder 10.

----------. 1992. “I Did Not Call Him ‘Fritz’: Personal Recollections of Professor F.A. Hayek.”

Constitutional Political Economy 3: 129-135.

Butos, William N. and Roger Koppl. 2003. “Science as a Spontaneous Order.” In The Evolution

of Scientific Knowledge, eds. H.S. Jensen, L.M. Richter, and M.T. Vendelø, 164–188.

Cheltenham: Edward Elgar.

Caldwell, Bruce. 2004. Hayek’s Challenge: An Intellectual Biography of F.A. Hayek. Chicago:

University of Chicago Press.

----------. 2010. “Introduction.” In The Collected Works of F.A. Hayek, Volume XV, The Market

and Other Orders, ed. B.J. Caldwell, 1–35. Chicago: The University of Chicago Press.

----------. 2014. “Introduction.” In The Collected Works of F.A. Hayek, Volume XIII, Studies on

the Abuse and Decline of Reason: Text and Documents, ed. B.J. Caldwell, 1–45.

Chicago: The University of Chicago Press.

Carnap, Rudolf, Hans Hahn and Otto Neurath. 1929. Wissenschaftliche Weltauffassung: Der

Wiener Kreis. Wien: Wolf. URL = http://againstpolitics.com/the-scientific-conception-of-

the-world-the-vienna-circle/

Duhem, Pierre. [1914] 1954. The Aim and Structure of Physical Theory, trans. P. W. Wiener.

Princeton, NJ: Princeton University Press.


45

Feyerabend, Paul. 1965. “Problems of Empiricism.” In Beyond the Edge of Certainty, ed. R.

Colodny, 145-260. New York: Prentice-Hall.

---------. [1975] 2010. Against Method. New York: Verso.

Hanson, Norwood R. 1958. Patterns of Discovery. New York: Cambridge University Press.

Hayek, F.A. 1952. The Sensory Order: An Inquiry into the Foundations of Theoretical

Psychology. Chicago: University of Chicago Press.

----------. 1967. Studies in Philosophy, Politics, and Economics. Chicago: University of Chicago

Press.

----------. [1956] 1967. “The Dilemma of Specialization.” In Studies in Philosophy, Politics, and

Economics, 122-132. Chicago: University of Chicago Press.

----------. [1967] 1992. “Ernst Mach (1838-1916) and the Social Sciences in Vienna.” In The

Collected Works of F.A. Hayek, Volume IV, The Fortunes of Liberalism: Essays on

Austrian Economics and the Ideal of Freedom, P.G. Klein, 172-175. London: Routledge.

---------. [1952] 2010. “Scientism and the Study of Society.” in The Collected Works of F.A.

Hayek, Volume XIII, Studies on the Abuse and Decline of Reason: Text and Documents,

ed. B.J. Caldwell, 75–166. Chicago: The University of Chicago Press.

----------. [1960] 2011. The Constitution of Liberty. In The Collected Works of F.A. Hayek,

Volume XVII, The Constitution of Liberty: The Definitive Edition, ed. R. Hamowy.

Chicago: University of Chicago Press.

----------. [1945] 2014. “The Use of Knowledge in Society.” In The Collected Works of F.A.

Hayek, Volume XV, The Market and Other Orders, ed. B.J. Caldwell, 93–104. Chicago:



46

----------. [1946] 2014. “The Meaning of Competition.” In The Collected Works of F.A. Hayek,

Volume XV, The Market and Other Orders, ed. B.J. Caldwell, 105–116. Chicago:


----------. [1955] 2014. “Degrees of Explanation.” In The Collected Works of F.A. Hayek, Volume

XV, The Market and Other Orders, ed. B.J. Caldwell, 195–212. Chicago: University of

Chicago Press.

---------. [1961] 2014. “A New Look at Economic Theory.” In The Collected Works of F.A.



----------. [1962] 2014. “The Economy, Science and Politics.” In The Collected Works of F.A.



----------. [1963a] 2014. “Rules, Perception and Intelligibility.” In The Collected Works of F.A.



----------. [1963b] 2014. “Economists and Philosophers.” In The Collected Works of F.A. Hayek,



----------. [1964a] 2014. “Kinds of Rationalism.” In The Collected Works of F.A. Hayek, Volume

XV, The Market and Other Orders, ed. B.J. Caldwell, 39–53. Chicago: University of

Chicago Press.


47

---------. [1964b] 2014. “The Theory of Complex Phenomena.” In The Collected Works of F.A.



----------. [1967a] 2014. “Notes on the Evolution of Systems of Rules of Conduct: The Interplay

between Rules of Individual Conduct and the Social Order of Actions.” In The Collected

Works of F.A. Hayek, Volume XV, The Market and Other Orders, ed. B.J. Caldwell, 278–

292. Chicago: University of Chicago Press.

----------. [1967b] 2014. “The Results of Human Action but Not of Human Design.” In The

Collected Works of F.A. Hayek, Volume XV, The Market and Other Orders, ed. B.J.

Caldwell, 293–303. Chicago: University of Chicago Press.

----------. [1968] 2014. “Competition as a Discovery Procedure.” In The Collected Works of F.A.



----------. [1970] 2014. “The Errors of Constructivism.” In The Collected Works of F.A. Hayek,



Hull, David L. 1988. Science as a Process: An Evolutionary Account of the Social and

Conceptual Development of Science. Chicago: University of Chicago Press.

Kuhn, Thomas S. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago

Press.

Lakatos, Imre. 1968–1969. “Criticism and the Methodology of Scientific Research

Programmes.” Proceedings of the Aristotelian Society 69: 149-186.


48

Latour, Bruno and Steve Woolgar. [1979] 1986. Laboratory Life: The Construction of Scientific

Facts, Second Edition. Princeton, NJ: Princeton University Press

Laudan, Larry. 1990. “Demystifying Underdetermination.” In Scientific Theories, ed. C. Wade

Savage, 267-297. Minneapolis: University of Minnesota Press.

Leonard, Thomas C. 2002. “Reflection on Rules in Science: An Invisible-Hand Perspective.”

Journal of Economic Methodology 9:2, 141-168.

Levy, David and Sandra Peart. Manuscript. “‘Almost Wholly Negative’: The Ford Foundation’s

Appraisal of the Virginia School” which includes the Buchanan-Gordon letter in question

McQuade, Thomas J. and William N. Butos. 2005. “The Sensory Order and Other Adaptive

Classifying Systems.” Journal of Bioeconomics 7: 335–358.

McQuade, Thomas J. 2007. “Science and Market as Adaptive Classifying Systems.” Advances in

Austrian Economics 9: 51-86.

----------. 2010. “Science and The Sensory Order.” Advances in Austrian Economics 13: 23-56.

Menger, Karl. 1994. Reminiscences of the Vienna Circle and the Mathematical Colloquium, eds.

L. Golland, B. McGuinness, and A. Sklar: Dordrecht: Kluwer Academic Publishers.

Polanyi, Michael. 1958. Personal Knowledge: Towards a Post-Critical Philosophy. Chicago:


----------. 1966. The Tacit Dimension. Garden City, NY: Doubleday & Company, Inc.

Popper, Karl R. 1984. “Evolutionary Epistemology.” In Evolutionary Theory: Paths into the

Future, ed. J.W. Pollard, 239-255. New York: Wiley.

Quine, W.V.O. [1951] 1961. “Two Dogmas of Empiricism.” In From a Logical Point of View,

2nd revised edition. New York: Harper and Row.


49

Ryle, Gilbert. 1946. “Knowing How and Knowing That.” Proceedings of the Aristotelian Society

46: 1-16.

Schmidtz, David. 2012. "Friedrich Hayek." In The Stanford Encyclopedia of Philosophy, ed.

E.N. Zalta, URL = http://plato.stanford.edu/archives/fall2012/entries/friedrich-hayek/

Smith, Barry. 1990. “On the Austrianness of Austrian Economics.” Critical Review 4: 212-238.

----------. 1994. Austrian Philosophy: The Legacy of Franz Brentano. Chicago: Open Court

Publishing Company.

Smith, Vernon L. 2008. Rationality in Economics. Cambridge: Cambridge University Press.

Suppe, Frederick. 1977. The Structure of Scientific Theories, 2nd ed., Urbana: University of

Illinois Press.

Toulmin, Stephen. 1953. Philosophy of Science: An Introduction. London: Hutchison’s

University Library.

----------. 1961. Foresight and Understanding. Bloomington, IN: University of Indiana Press.

----------. 1981. “Evolution, Adaptation, and Human Understanding.” In Scientific Inquiry and

the Social Sciences, eds. M.B. Brewer and B.E. Collins, 18–36. San Francisco: Jossey-

Bass.

Weimer, Walter B. 1980. “For and Against Method: Reflections on Feyerabend and the Foibles

of Philosophy.” Pre/Text 1-2: 161-203.